Hour graph points not accurate and duplicating

Below is the output on my script vs dashboard graph values. Seems like the graphs are trying to average? The last data point on the hour graph for Temperature reads 75.20 from the sensor but shows up as 75.18 on the graph (minor difference but still different). The rest of the values are also just slightly off. You can also see 2 data points close to each other, but I only ever send 1 value at a time so I’m not sure where the other ones comes from.

[1264] Connected to WiFi
[1264] My IP: 10.0.0.150
[5001] Connecting to arduino.mydevices.com:8442
[5388] Ready (ping: 155ms).
Entered Temerature
Temerature: 73.40
Entered Humidity
Humiditdy: 41.00
Entered Heat Index
Heat Index: 72.27
Got all values - sleeping
[1262] Connected to WiFi
[1262] My IP: 10.0.0.150
[5001] Connecting to arduino.mydevices.com:8442
[5400] Ready (ping: 157ms).
Entered Temerature
Temerature: 73.40
Entered Humidity
Humiditdy: 41.00
Entered Heat Index
Heat Index: 74.25
Got all values - sleeping
[3259] Connected to WiFi
[3259] My IP: 10.0.0.150
[5001] Connecting to arduino.mydevices.com:8442
[5348] Ready (ping: 161ms).
Entered Temerature
Temerature: 75.20
Entered Humidity
Humiditdy: 40.00
Entered Heat Index
Heat Index: 72.23
Got all values - sleeping
[1262] Connected to WiFi
[1263] My IP: 10.0.0.150
[5001] Connecting to arduino.mydevices.com:8442
[5377] Ready (ping: 162ms).
Entered Temerature
Temerature: 75.20
Entered Humidity
Humiditdy: 41.00
Entered Heat Index
Heat Index: 72.13
Got all values - sleeping

Looks like the double data points are fixed now.

Thinking about this some more I’m not sure what the best way to handle the actual values displayed would be. Either you select a point in time that has a real value or you have to average to get close. I guess since my ESP is only updating every 10 minutes that’s why I’m noticing it. In reality the hour graph is my live graph. Maybe you could have different options for displaying data on the graph? Ex. actual values at a certain time or averaged values (possibly even mean/median/mode options) but in my case I would need the actual values to search until it actually finds a value. Seems like something that will have to be discussed to determine what is feasible.